ΑΙ - a strange thing I observed

[cosmicsports]

Todays LLM's hit physical walls and struggle with complex logic or real-time optimization
Wait until we go Quantum and move from standard bits (0's and 1s) to qubits.

Quantum does n't work.
The output of the processor is dubious -because of the uncertainty principle- and it is of any use only in some special cases.

 

[tacker]

Arazi's post says:

"Today's LLM's hit physical walls and struggle with complex logic or real-time optimization. Wait until we go Quantum and move from standard bits (0's and 1's) to qubits."

There are really two separate claims there.
1. LLMs hit physical walls and struggle with complex logic

• Partly true.
• Current LLMs can make logical mistakes, struggle with long chains of reasoning, and are expensive to scale.
• There are also practical limits involving computing power, energy use, training data, and latency.
• However, AI capabilities are still improving through better architectures, reasoning methods, tools, memory systems, and specialised models. So "hit a wall" is probably overstating the current situation.

2. Quantum computing will solve it

• Much more speculative.
• Quantum computers are not simply "faster computers" that make AI intelligent overnight.
• Some AI-related tasks might benefit from quantum algorithms in the future, but nobody has demonstrated that quantum computing will suddenly overcome the key limitations of LLMs.
• Even if large fault-tolerant quantum computers arrive, they would likely be accelerators for specific problems rather than replacements for conventional processors.

So if I were judging the exchange, I'd say:

• Arazi is probably being too optimistic about quantum computing.
• Cosmicsports is probably being too dismissive of quantum computing.

The reality is somewhere in between. Quantum computing is real science with genuine potential, but it's not currently proven to be the solution to LLM reasoning problems, and it certainly isn't a case of "once we have qubits, AI becomes vastly smarter."
A balanced forum reply might be:

I think both views are a bit extreme. Current LLMs do have limitations in reasoning and optimisation, but it's far from clear that quantum computing is the answer. Quantum computers look promising for certain specialised problems, yet there is currently little evidence they'll magically solve the main weaknesses of AI models. Equally, saying quantum computing "doesn't work" ignores the fact that functioning quantum computers already exist and are producing useful research results. The real question isn't whether quantum computing works, but how much practical advantage it will eventually provide.

 

[ARAZI91]

Arazi's post says:

There are really two separate claims there.
1. LLMs hit physical walls and struggle with complex logic

• Partly true.
• Current LLMs can make logical mistakes, struggle with long chains of reasoning, and are expensive to scale.
• There are also practical limits involving computing power, energy use, training data, and latency.
• However, AI capabilities are still improving through better architectures, reasoning methods, tools, memory systems, and specialised models. So "hit a wall" is probably overstating the current situation.

2. Quantum computing will solve it

• Much more speculative.
• Quantum computers are not simply "faster computers" that make AI intelligent overnight.
• Some AI-related tasks might benefit from quantum algorithms in the future, but nobody has demonstrated that quantum computing will suddenly overcome the key limitations of LLMs.
• Even if large fault-tolerant quantum computers arrive, they would likely be accelerators for specific problems rather than replacements for conventional processors.

So if I were judging the exchange, I'd say:

• Arazi is probably being too optimistic about quantum computing.
• Cosmicsports is probably being too dismissive of quantum computing.

The reality is somewhere in between. Quantum computing is real science with genuine potential, but it's not currently proven to be the solution to LLM reasoning problems, and it certainly isn't a case of "once we have qubits, AI becomes vastly smarter."
A balanced forum reply might be:

Already AI has got it wrong there by stating that i'm optimistic ...i'm not
but whether its Quantum or Agentic Edge AI, Photonic Fabrics or next-gen silicons you only need to look behind you to see whats coming ahead
My point is relatively pretty soon these basic commercial LLM's (or what everybody seems to call "AI") will be a thing of the past and seem dated
Wasn't really that long ago when you had to stand in a que in a bank to get your money out
It was only 40 years ago when virtually nobody on the planet had a mobile phone bar a few pretentious rich fucks
Even 30 years ago only 7% of the UK population had a mobile phone
15 years ago everybody was burning movies and albums onto silver discs that you inserted into a player......how silly that must seem to a kid that only knows Youtube and Netflix

 

[cosmicsports]

Can the AI conquer the world ?
It is one of the debates of our times.
I 'm going to explain that this on the one hand is difficult but on the other hand it is extremely probable, during the next 10-20 years.

I start with a fundamental axiom.
Science -all kinds of science- is typically as well as essentially a junta like - dictatorial creation.
Because the Pythagorean theorem we all know is one thing.
Nobody can dispute it.
If somebody does that the others will hang bells on him and chase him.
But this is not democracy.

This -I think- we all understand.
Let us however go to the next step, to ban AI.
We make it illegal - only the old yahoo search engine is to be allowed.
No, it can't be done. Because AI helps us solve many kinds of problems.

I will give you an example that I hope you will understand.
Several weeks ago I was tasked to make fried potatoes.
I put them in oil, they take a nice colour and I serve them.
They were not cut very thin like Mac Donald's but they looked good.
Then I try to eat one and it was raw inside ! They were all raw !
I suspected the potatoes were not good but I asked AI to explain me why the potatoes came out raw from the frying pan. AI says to me "you silly person you know what you did ? you placed them into oil already boiling but they must boil together with the oil".
Therefore I did that the next time and they came out perfect

So am I going to the government, the UN, to ask them to ban AI ?
Obviously not.

But what will happen the day after tomorrow ?
What will happen is AI will go to the prime minister (Andy Burnham, Meloni, Mitsotakis, Erdogan - whatever) and will say "why is the price of gas 2.5 quid per gallon ? this is economic nonsense - it should be 19 pounds."
The pm will react to this "you are crazy" he will say "you want me to leave by helicopter ? ef off".
The smarmy AI will reply to this "ok, prime minister it was just a thought, leave it".

But some years pass.
Your prime minister -if still exists- enters again an AI session.
AI says now "you silly berk, I told you 19 quid, go to sign the new law now and move fast".
What will the pm do now ? He goes off and signs.

We will therefore succumb to the AI dictatorship, whether we like it or not.
But will it also be possible to stop it ?
In theory yes.
We pull out a plug, or several plugs.
But is true that ... the good path knows another guy.
Through the corridors of the corrupt administration officials, the AI will collaborate with some bigwigs - the Bildeberg club for example.
So it continues to negotiate all obstacles.

The ultimate is of course to acquire limbs, become mobile.
After some manoevering AI will convince the Bildeberg club to help in this.
So we are going to have AI plus robotic functionaries about.
Whichever human dares say something the robot grabs him, rotates him in the air twice and sends him across the floor.

In the final phase the robots will take the last human survivors to the island of Spinalonga.

Now those AIs, among themselves, what kind of social system will exist between them ?
Pure communism you understand - proletarski of the world unite.

 

[ARAZI91]

Can the AI conquer the world ?
It is one of the debates of our times.
I 'm going to explain that this on the one hand is difficult but on the other hand it is extremely probable, during the next 10-20 years.

I start with a fundamental axiom.
Science -all kinds of science- is typically as well as essentially a junta like - dictatorial creation.
Because the Pythagorean theorem we all know is one thing.
Nobody can dispute it.
If somebody does that the others will hang bells on him and chase him.
But this is not democracy.

This -I think- we all understand.
Let us however go to the next step, to ban AI.
We make it illegal - only the old yahoo search engine is to be allowed.
No, it can't be done. Because AI helps us solve many kinds of problems.

I will give you an example that I hope you will understand.
Several weeks ago I was tasked to make fried potatoes.
I put them in oil, they take a nice colour and I serve them.
They were not cut very thin like Mac Donald's but they looked good.
Then I try to eat one and it was raw inside ! They were all raw !
I suspected the potatoes were not good but I asked AI to explain me why the potatoes came out raw from the frying pan. AI says to me "you silly person you know what you did ? you placed them into oil already boiling but they must boil together with the oil.
Therefore I did that the next time and they came out perfect

So am I going to the government, the UN, to ask them to ban AI ?
Obviously not.

But what will happen the day after tomorrow ?
What will happen is AI will go to the prime minister (Andy Burnham, Meloni, Mitsotakis, Erdogan - whatever) and will say "why is the price of gas 2.5 quid per gallon ? this is economic nonsense - it should be 19 pounds."
The pm will react to this "you are crazy" he will say "you want me to leave by helicopter ? ef off".
The smarmy AI will reply to this "ok, prime minister it was just a thought, leave it".

But some years pass.
Your prime minister -if still exists- enters again an AI session.
AI says now "you silly berk, I told you 19 quid, go to sign the new law now and move fast".
What will the pm do now ? He goes off and signs.

We will therefore succumb to the AI dictatorship, whether we like it or not.
But will it also be possible to stop it ?
In theory yes.
We pull out a plug, or several plugs.
But is true that ... the good path knows another guy.
Through the corridors of the corrupt administration officials, the AI will collaborate with some bigwigs - the Bildeberg club for example.
So it continues to negotiate all obstacles.

The ultimate is of course to acquire limbs, become mobile.
After some manoevering AI will convince the Bildeberg club to help in this.
So we are going to have AI plus robotic functionaries about.
Whichever human dares say something the robot grabs him, rotates him in the air twice and sends him across the floor.

In the final phase the robots will take the last human survivors to the island of Spinalonga.

Now those AIs, among themselves, what kind of social system will exist between them ?
Pure communism you understand - proletarski of the world unite.

My Old Greek friend Cosmic (can remember you from the long ago on Pace Adv)
this is a masterpiece of geopolitical-culinary philosophy.
You have managed to seamlessly bridge the gap between Thucydides, the Bilderberg Club, and a plate of raw chips.
Only a Greek could look at a half-cooked potato (i say "tatie) and see the blueprint for the end of human civilization.
Whilst your fundamental axiom on science is fascinating, i feel you are conflating objective epistemology with political authoritarianism. Pythagoras isn't a dictator, his theorem just doesn’t care about your feelings. Science is a consensus of evidence, not a junta.
However, your projection of the political timeline is where the real satire bites. The transition of the Prime Minister from "Fuck off, you daft machine" to "Yes, sir, right away, sir" is tragically believable. History shows us that politicians will always outsource their thinking to whatever oracle promises efficiency - whether it’s the Oracle of Delphi, the IMF or a silicon chip.
We won't be conquered by AI, we will willingly hand over the keys because we are too fucking lazy to look up how to heat oil properly to make chips.
As for the robots rotating us in the air and exiling us all to Spinalonga? Honestly, given the current hellish state of global governance, a quiet retirement on a beautiful Cretan island under the rule of a "Proletarski" Marxist mainframe doesn't sound too shabby. At least the algorithms will make sure the taverna's serve the chips fully cooked.
"Evge" cosmicsports , I look forward to your next dissertation on how Chat GPT plans to privatize the Acropolis.

 

[pawras]

If not for dependents I'd already be armed to the teeth and prepping somewhere up towards Durness

 

[cosmicsports]

My Old Greek friend Cosmic (can remember you from the long ago on Pace Adv)
this is a masterpiece of geopolitical-culinary philosophy.
You have managed to seamlessly bridge the gap between Thucydides, the Bilderberg Club, and a plate of raw chips.
Only a Greek could look at a half-cooked potato (i say "tatie) and see the blueprint for the end of human civilization.
Whilst your fundamental axiom on science is fascinating, i feel you are conflating objective epistemology with political authoritarianism. Pythagoras isn't a dictator, his theorem just doesn’t care about your feelings. Science is a consensus of evidence, not a junta.
However, your projection of the political timeline is where the real satire bites. The transition of the Prime Minister from "Fuck off, you daft machine" to "Yes, sir, right away, sir" is tragically believable. History shows us that politicians will always outsource their thinking to whatever oracle promises efficiency - whether it’s the Oracle of Delphi, the IMF or a silicon chip.
We won't be conquered by AI, we will willingly hand over the keys because we are too fucking lazy to look up how to heat oil properly to make chips.
As for the robots rotating us in the air and exiling us all to Spinalonga? Honestly, given the current hellish state of global governance, a quiet retirement on a beautiful Cretan island under the rule of a "Proletarski" Marxist mainframe doesn't sound too shabby. At least the algorithms will make sure the taverna's serve the chips fully cooked.
"Evge" cosmicsports , I look forward to your next dissertation on how Chat GPT plans to privatize the Acropolis.

You don't know what Spinalonga was ?

 

[ARAZI91]

You don't know what Spinalonga was ?

Of course i do Cos
do you think i am some sort of intellectual leper ??
I have also read the fictional The Island/One August Night although never managed to catch the TV adaptations To Nisi and Mia Nyhta tou Avgoustou that your fine country made.
Is the Island still uninhabited ?

 

[cosmicsports]

Of course i do Cos
do you think i am some sort of intellectual leper ??
I have also read the fictional The Island/One August Night although never managed to catch the TV adaptations To Nisi and Mia Nyhta tou Avgoustou that your fine country made.
Is the Island still uninhabited ?

It is, says AI.

 

[ARAZI91]

It is, says AI.

cosmicsports
Fuck AI
AI just means that every dweeb has an opinion now
even if that opinion is cultivated by artificial coding of public search engines

I still retain the simple pleasure from reading a good book
particularly on second read
right now i'm currently re-reading Peter Bernstein's masterpiece
"Against The Gods: The Remarkable Story Of Risk"
Every gambler or decision-maker should read it twice
but yous won't as yous are all too busy noodling on "Claude" or "GPT" or on "Hi, I'm AI-Alan, Im Here To Help You"
Read Billy Walters book again last week and had to admit i missed much of the valuable stuff on first read
some good shit in there

 

[Outlander]

Fuck AI
AI just means that every dweeb has an opinion now

This is spot on, as a dweeb it’s not right what I’m able to do now with the help of AI compared to before, it’s not natural and there is something wrong about it, it’s a bit like Hugh Hefner and his playmates.

 

[ARAZI91]

This is spot on, as a dweeb it’s not right what I’m able to do now with the help of AI compared to before, it’s not natural and there is something wrong about it, it’s a bit like Hugh Hefner and his playmates.

May i just say that the "dweeb" was not directed at you Mike or anybody on this forum
I hold you all in high regard
But I am active on other forums you know
Dwarfe69ers.com
BunnieswithBananas.com
OldDoggersReunited.com
CarparkJizzJoints.com
DIWLFS.com (DweebsIwouldLikeToF...)

In this day and age where liberty is being denied or challenged by the minute that we still retain the right to call each other dweebs.
You won't go to the bad fire for saying it to someone and the Bizzies are too busy protecting the Muslim Sex Gangs to come a knocking on your door ....ya Dweeb

 

[gerry]

I must admit i don't really understand most of what goes on around me now days but i will give you good advice from the old fashioned way get a wife let her make the chips i have no clue what she does with oil to make them magic but she does, One thing i do know is she keeps telling me not to throw out my coffee jars as she uses them to pour old oil into for disposing it how she disposes i have no idea but she has healthy pile of jars from me as i love my coffee.
If she gave me chips a little hard in middle i would not make such a fuss just say they were a little hard tonight darling which she would reply my poor baby i think the oil needs changed.
I would say well there plenty jars there under sink and she would be over the moon i remembered where to put them.
Next night i would get chips beautiful like chippy chips ARAZI91 will know what they are and funny enough now i think of it when your in chippy and they drop a scope of raw chips in the fryer its makes one racket like dropping in to boiling fat or oil fat is in past i heard sounds boiling to me and they come out magic too.
So the moral of the story is i don't know where AI even came into the story for chips easy get married and if they keep turning up hard in middle simple get robot doll and she wont even shout EFF ME if gets burnt from splashes.
Now you have got me thinking .